A Robust Orbit Determination Algorithm for Space-Based Optical Observations in Complex Environments

With the rapid growth of space targets and the urgent need for near-Earth asteroid monitoring, as well as the increasing demands of deep-space exploration missions, accurate orbit determination is essential for space safety and reliable navigation. However, traditional least squares methods lack robustness under outlier data and large initial errors, which can lead to estimation failure. To address these challenges, this paper proposes a robust orbit determination algorithm for space-based optical observations in complex environments. The algorithm reconstructs the nonlinear observation equation into a cross-product form of the line-of-sight vector, establishing a linear measurement model, and adopts an optimization strategy that minimizes the sum of the absolute values of residuals. By introducing slack variables, the orbit determination problem is transformed into a linear programming problem, achieving automatic outlier rejection. Furthermore, the analytical covariance at the estimated epoch was also derived. Simulation results show that the proposed algorithm effectively mitigates the impact of outliers and enhances convergence, enabling more robust and accurate orbit determination in complex observation environments.